Robert Frank’s Strange Case for Taxing ‘The Rich’

November/​December 2007 • Policy Report
By David R. Henderson

In testimony before the House Financial Services Committee on May 16 of this year, Cornell University economist Robert H. Frank gave a new justification for a progressive consumption tax — that is, a tax on consumption featuring higher tax rates for those who consume more. He argued that it’s good for them. His testimony is the culmination of a line of thinking that Frank has been developing for many years, starting with journal articles in the early 1980s that led to his 1985 book, Choosing the Right Pond. Because of his high status in the economics profession and because his views on tax policy and other government policy seem to have struck a chord, it is worth examining those views. Frank’s justifications for higher tax rates on people with higher incomes rest on a very slender reed, and in making his case for them he ends up contradicting, in two ways, the very foundations of his views.

A summary of Frank’s argument is in order. He claims that many of the goods we buy are “positional.” In other words, their value to those who own them depends strongly on relative position rather than anything absolute. Frank gives the example of a Ferrari Scaglietti, a car that sells in the United States for about $250,000. According to Frank, purchases of such cars and of 60,000-square-foot houses “subtly change the social frame of reference that defines what kinds of houses and cars seem necessary or appropriate.” The people who buy such things up the ante on their purchases, and then the people “below” them do likewise, and so on down the income scale. Frank calls this alleged phenomenon an “expenditure cascade.” In buying positional goods, the highest‐​income people, writes Frank, impose a negative externality on the people below them, who then, through their purchases, impose a negative externality on those below them, and so on. Frank advocates the standard economist’s solution to a negative externality, which is a tax on the activity that generates the externality. Frank’s favored tax is a tax on consumption, with a higher rate for those who consume more.

As a bonus, argues Frank, a government can tax high‐​income people even more than it currently does without making them worse off. How so? For simplicity, imagine a society in which there are a million people making more than $500,000 a year. Most of us would agree, I think, that those people have high incomes. Imagine that they now pay 30 percent of their income in federal income taxes. Now imagine that the government, following Frank’s suggestion, imposes a tax on consumption above some amount per year and, thus, raises tax rates on high‐​income people so that those million people now pay 40 percent of their income in federal income taxes. Because their relative position with respect to each other would be unchanged, and because they spend so much money on positional goods anyway, they would not care — or so the argument goes. As Frank testified, “Thus, if a consumption tax led wealthy families to buy 5,000-square-foot houses instead [of] 8,000, and Porsche Boxsters instead of Ferraris, no one would really be worse off, and several hundred thousand dollars of resources per family would be freed up for more pressing purposes.” The government could then take the extra revenue generated by the higher tax rates and spend it on things that people, including many of those with high incomes, value. Because the added tax has a zero cost to those taxed and the revenues create benefits for at least some members of society, the tax creates net benefits. That is, in a nutshell, Frank’s argument for higher taxes on people with high incomes.

Frank adopts Fred Hirsch’s characterization of positional goods: “goods that are sought after less because of any absolute property they possess than because they compare favorably with others in their class.” Frank also writes, “Positional goods are, by their nature, things in fixed supply.” He gives houses, cars, and jobs as examples of positional goods and medical care and leisure as examples of nonpositional goods. And yet, his examples seem to belie his definition. While it’s true that certain jobs — chairman of Microsoft, for example — are in fixed supply, houses and cars are not. And yet, in Frank’s mind, they’re positional. Leisure, on the other hand, if it involves courtside seats at a New York Knicks game, seems to involve fixed supply: the number of such seats is strictly limited. Given how important positional goods are to Frank’s whole scheme, it’s surprising how he doesn’t seem to follow his own definition in classifying goods one way or the other while still seeming to be quite confident about which is which.

Ignoring the Changing Evidence

A pillar of Frank’s argument is that a large percentage of people care about their relative position. In Choosing the Right Pond, he defends that assumption by pointing to anomalies in the pay structure of various firms, anomalies that he attributes to people caring about relative position. Most of his anomalies have to do with pay structures that, Frank argues, are “flatter” than standard economics would predict. Standard economics states that workers are paid an amount roughly equal to the value of their marginal product — that is, the increment in value that is due to their being in the firm. But, notes Frank, if this were true, one would expect to see great disparities between the salaries of workers who have great differences in productivity. He points to, among other things, the University of Michigan pay scale for economists in 1983–84, where the highest salary was only a little more than double the lowest. He never mentions the fact that the University of Michigan is a government bureaucracy, making it not the best test of the standard economics account of freemarket wages. Nor does he mention that one of the main ways the stars of academic institutions are “paid” is with lower teaching loads and more research funds.

Even more interesting is how the world seems to have changed since Frank began writing about these issues and the contortions he goes through to sustain his argument for higher taxes. When he first began, he argued that relatively flat pay structures are indirect evidence for his view that people care a lot about relative position. But in his May 2007 testimony, Frank noted that the “anti‐​raiding norms of business have recently begun to unravel” so that, now, pay for top managers can be a huge multiple of pay for bottom managers. In other words, it would seem, many top managers are being paid an amount that approximates their marginal product. You might think that this would cause Frank to reexamine his earlier strongly held views. But he doesn’t.

Instead, he comes up with a new argument for progressive consumption taxes. He now argues that too many people are vying for the top jobs because of the higher pay those jobs carry. They are fighting, he argues, over a fixed pie and, in a variant of the famous “tragedy of the commons,” he compares the competition for the top jobs to gold prospecting. He testified that “the gold found by a newcomer to a crowded gold field is largely gold that would otherwise have been found by others.” Similarly, he argues, “an increase in the number of aspiring hedge fund managers produces much less than a proportional increase in the amount of commissions on managed investments.”

But he can’t hold on to this argument for even a page. Just four paragraphs later, he testified: “A slightly more talented CEO or hedge fund manager can boost a large organization’s annual bottom line by hundreds of millions of dollars or more.” Exactly. It does make sense, therefore, for companies to look for small differences in talent because those differences can cause huge increases in profits. The problem with Frank’s tragedy of the commons analogy is that there is no commons. The tragedy of the commons occurs when no one owns the resource: thus the word “commons.” But those who hire hedgefund managers own their resources, so one would not expect overinvestment in being the manager. Frank implicitly admits this, writing, “To be sure, even those who fail to win the biggest prizes often go on to earn comfortable incomes.” But in the very next sentence, he retreats to his old position, saying, “But career choices must be measured not in terms of absolute pay but relative to what might have been” (emphasis added). This is astounding. More than 20 years ago, Frank argued, as an empirical matter, that people care about relative income. Now in the face of evidence that absolute income matters a lot to them — otherwise, why would anti‐​raiding norms have unraveled — he argues that it shouldn’t — thus his use of the word “must.” If the people don’t conform to his assumptions, it seems, we should tell them to.

I may have interpreted Frank’s use of the word “must” incorrectly because he goes on to write, “Contestants for the top prizes in finance are highly talented people who could have held interesting jobs at high pay in other fields. Those who end up as account managers in small banks may not starve, but neither do they realize their full potential.” Frank’s argument here, presumably, is that people overestimate their expected return from competing for superstar jobs and so they overinvest in competing for them. “Overinvest” is determined relative to a baseline of efficient allocation of people to jobs. This is what we “must” compare the actual outcome to. But this is incredibly presumptuous on his part. Does Frank really think tax policy ought to encourage would‐​be novelists to go to medical school? He writes as if the world is a place of certainty and he heavily discounts the extent to which competition is a discovery process.

Taxes, Work, and GDP

Interesting also is how Frank deals with the supply‐​side economists’ argument that higher marginal tax rates reduce effort, and how his argument has evolved. In Choosing the Right Pond, Frank accepted the view that higher marginal tax rates do, indeed, reduce work effort — and applauded that result. Frank wrote:

The real problem is not at all that the current tax system induces people to work too little, take too few risks, and so on. On the contrary, it is a lack of taxation that would cause individually rational citizens to work too many hours, take too many risks, and spend too little time with family and friends (emphasis in original).

By the time his 1999 book, Luxury Fever, was published and in his 2007 testimony, though, Frank had changed his argument. Interestingly, while he correctly used the term “supply‐​sider” in his 1985 book, by 1999 he no longer used that term; instead he used the disparaging term “trickle‐​down theory” to label the supply‐​side theory that changes in marginal tax rates affect economic behavior. (No supply‐​sider calls himself a trickle‐​down theorist: this is the term used exclusively by critics of supplyside economists. Frank’s use of the term “trickle‐​down” suggests bad faith on his part.) Interestingly, Frank now argues that marginal tax rates do not clearly reduce work effort and briefly dismisses the substantial evidence that supply‐​side economists such as Harvard’s Martin Feldstein have presented. Nowhere could I find Frank acknowledging this complete reversal of his 1985 argument. Was he wrong then or is he wrong now?

Furthermore, in arguing for a progressive consumption tax, Frank contradicts another big part of his earlier work without ever acknowledging it: he argues that a progressive consumption tax will increase GDP. That is doubtful, but let’s accept it for a minute. Why is this good? I know why I think it’s good: GDP is a rough measure of human welfare. Of course, there are huge problems with GDP as a measure of welfare, two of the most important being GDP’s failure to value leisure time and its valuing of government expenditures at cost. But we can put those aside because they do not relate to Frank’s argument. Why does Frank think GDP is a good measure of welfare? After all, he has spent a large part of the quarter century arguing that it is not a good measure. In Luxury Fever, he recalls his time in the Peace Corps in Nepal:

My one‐​room house had no electricity, no heat, no indoor toilet, no running water. The local diet offered little variety and virtually no meat. Yet, although my living conditions in Nepal were a bit startling at first, the most salient feature of my experience there was how quickly they came to seem normal. Within a matter of weeks, I lost all sense of impoverishment. Indeed, my $40 monthly stipend was more than most others had in my village, and with it I experienced a feeling of prosperity that I have recaptured only in recent years.

I don’t doubt any of this. Indeed, my guess is that this experience heavily influenced his view that what matters is relative income because that’s what mattered to Frank: notice how, even in Nepal, Frank felt the need to compare himself with those around him. But this raises two questions. First, why didn’t he move back to Nepal? Why did he spend those intervening decades between the late 1960s and the late 1990s, when he finally recaptured the feeling of prosperity, in the United States? Second, if huge components of GDP — meat, indoor plumbing, electricity — don’t matter much, how can he justify his policies by arguing that they increase U.S. GDP? Shouldn’t he want to take, not just a huge slice of the highest‐ income people’s income, but also a substantial slice out of every American’s income and, say, give it to people in poor countries? His failure to advocate this is certainly not because Frank has qualms about forcibly taking money from people. In fact, he even labels (in Econ Journal Watch) as “crybabies” those who object that taxation is coercive. To his credit, Frank has argued that real GDP does matter because it allows us to help more poor people and to extend our lives with medical technology. But it’s hard to know why he draws the line on government wealth distribution policies at the U.S. border.

Frank argues that consumption taxes on higher‐​income people make them no worse off because, as noted, they care about relative income, not absolute income. And, presumably, these people put at least a tiny positive value on the things government would spend the additional revenues on. Is Frank open to testing his assumption? Here’s a test. If he’s right, a majority of those high‐​income people, indeed a supermajority, would vote in favor of higher taxes on themselves. Frank should propose a referendum on whether to raise tax rates on highincome people, with only high‐​income people able to vote. If he is right, such a referendum would pass overwhelmingly. I predict that such a referendum would go down in flaming defeat. If I’m right, then the whole empirical basis of his argument is wrong.

Why We Want Things

It is true that we often want something when we see that someone else has it. But what doesn’t ring true is Frank’s view about why we want things. One of my earliest instances of an intense want was in 1955, when the coonskin cap came along after Walt Disney had made Davy Crockett famous. I saw some of my friends wearing them and I badly wanted one. My father, though, would not buy one for my brother or me. I remember the intense pain I had about not having it. But did I want that coonskin cap because I was competing with my friends for status and position? Not at all. I wanted it because it was so neat. Now, you might doubt the memory of a 56‐​year‐​old about his introspection 52 years earlier. Fine. Then consider this case. I also remember when the Ford Mustang and the Mercury Park Lane came along in 1965, when I was 14. I wanted either one of those cars badly. I tore out the full‐​page magazine ads picturing those cars, taped them on my wall, and pined for them every day. But the reason I wanted them was not that I saw people around me with them. I lived in a small town in rural Canada where you didn’t see new cars as soon as they came out. I had seen the ads for these cars and started yearning for them long before anyone in my town owned one. So, why did I want one of these cars? Because they were just so beautiful. I’ve asked other friends why they want the new expensive gadgets when they come out and invariably the answer is that they’re such neat toys. Few mention that they want them because they want to be higher up on the positional scale.

The closest Frank came in his testimony to giving direct evidence for his positional goods hypothesis was his evidence on housing. He noted that between 1980 and 2001, the median size of a new house increased from 1,600 square feet to more than 2,100 square feet, while the median family’s real income increased by less than 15 percent. This small increase in real income, he asserts, is “not nearly enough to comfortably finance so much larger a house.” Oh, really? Has Frank checked mortgage interest rates in 1980 versus 2001? In 1980, they averaged 12.7 percent; by 2001, they had fallen to 7 percent. This 45‐​percent drop in interest rates certainly did help people buy houses that were 31‐​percent bigger.

It’s true that the majority of things we want in our lives are things we want because we see other people with them. But Occam’s Razor applies here. The most straightforward reason is that when we see others with them, we see how those things might improve our lives. Think about some of the innovations that have come along just in my lifetime. I remember when Milton Friedman had open‐​heart surgery in December 1972. That was a fairly new surgery. But when that surgery worked for him and hundreds of other people, thousands of people could see how open‐​heart surgery would improve their lives — and they got such surgery, too. I would never have wanted a cell phone if I hadn’t seen others using them, because I wouldn’t have known they existed. But now that I do know, I see their value in making my life better, such as the time I landed at an unfamiliar airport and rendezvoused with a friend who was picking me up.

Interestingly, for someone who gets credit for thinking broadly about socioeconomic issues, Frank actually thinks quite narrowly about them. First, he tends to think that everyone is like him in having a strong comparative impulse. But this is false. (Abraham Buunk et al., the Journal of Social and Personal Relationships, found this in some people but not others.) And although Frank sometimes admits that one’s concern with relative social standing will rear its ugly head in matters not just of relative consumption — how about the following: “I spent my leisure time better than you did” — he always jumps back to assuming, without much evidence, that the dominant form of statusseeking is narrowly economic.

And beyond all that, what if Frank were completely correct in his assertion that many or most people care about relative income and position? I don’t doubt that some people are that way. My own solution is not to have such people as friends. But how would that justify forcibly taking their money? Wouldn’t the proper thing be to persuade people not to care about others’ income and even to work on one’s own psyche rather than to force one’s views on others? Frank’s advocacy of higher taxes reminds me of a scene from the TV show Scrubs. Carla, a nurse on the show, suggests to the janitor that they collect money from other employees to do a good deed.

JANITOR: I’ll check their lockers.
CARLA: I meant ask them.
JANITOR (with a quizzical look on his face): That seems kind of roundabout, but OK.

At one time, critics of economic freedom justified high taxes on high‐​income people on the grounds of ability to pay. They at least admitted that those taxes hurt those people. But the growing availability to even the poor of goods that were only recently thought of as luxury goods has weakened that argument. Now, Robert Frank argues for higher taxes on high‐​income people on the grounds that it is good for them. If that is the best the proponents of higher taxes come up with, maybe we should see this as intellectual progress.

Download the Policy Report Article

About the Author
David R. Henderson

Research Fellow, Hoover Institute